33
31
from xml.etree.cElementTree import (ElementTree, SubElement, Element,
34
32
XMLTreeBuilder, fromstring, tostring)
35
33
import xml.etree as elementtree
36
# Also import ElementTree module so monkey-patching below always works
37
import xml.etree.ElementTree
38
34
except ImportError:
39
35
from cElementTree import (ElementTree, SubElement, Element,
40
36
XMLTreeBuilder, fromstring, tostring)
41
import elementtree.ElementTree
42
38
ParseError = SyntaxError
43
39
except ImportError:
44
40
mutter('WARNING: using slower ElementTree; consider installing cElementTree'
53
49
from bzrlib import errors
56
class XMLSerializer(Serializer):
57
"""Abstract XML object serialize/deserialize"""
59
squashes_xml_invalid_characters = True
61
def read_inventory_from_string(self, xml_string, revision_id=None,
62
entry_cache=None, return_from_cache=False):
52
class Serializer(object):
53
"""Abstract object serialize/deserialize"""
55
def write_inventory(self, inv, f):
56
"""Write inventory to a file"""
57
raise NotImplementedError(self.write_inventory)
59
def write_inventory_to_string(self, inv):
60
raise NotImplementedError(self.write_inventory_to_string)
62
def read_inventory_from_string(self, xml_string, revision_id=None):
63
63
"""Read xml_string into an inventory object.
65
65
:param xml_string: The xml to read.
66
66
:param revision_id: If not-None, the expected revision id of the
67
67
inventory. Some serialisers use this to set the results' root
68
revision. This should be supplied for deserialising all
69
from-repository inventories so that xml5 inventories that were
70
serialised without a revision identifier can be given the right
71
revision id (but not for working tree inventories where users can
72
edit the data without triggering checksum errors or anything).
73
:param entry_cache: An optional cache of InventoryEntry objects. If
74
supplied we will look up entries via (file_id, revision_id) which
75
should map to a valid InventoryEntry (File/Directory/etc) object.
76
:param return_from_cache: Return entries directly from the cache,
77
rather than copying them first. This is only safe if the caller
78
promises not to mutate the returned inventory entries, but it can
79
make some operations significantly faster.
82
return self._unpack_inventory(fromstring(xml_string), revision_id,
83
entry_cache=entry_cache,
84
return_from_cache=return_from_cache)
71
return self._unpack_inventory(fromstring(xml_string), revision_id)
85
72
except ParseError, e:
86
73
raise errors.UnexpectedInventoryFormat(e)
88
75
def read_inventory(self, f, revision_id=None):
91
return self._unpack_inventory(self._read_element(f),
77
return self._unpack_inventory(self._read_element(f),
95
79
except ParseError, e:
96
80
raise errors.UnexpectedInventoryFormat(e)
115
99
return ElementTree().parse(f)
118
def escape_invalid_chars(message):
119
"""Escape the XML-invalid characters in a commit message.
121
:param message: Commit message to escape
122
:return: tuple with escaped message and number of characters escaped
126
# Python strings can include characters that can't be
127
# represented in well-formed XML; escape characters that
128
# aren't listed in the XML specification
129
# (http://www.w3.org/TR/REC-xml/#NT-Char).
130
return re.subn(u'[^\x09\x0A\x0D\u0020-\uD7FF\uE000-\uFFFD]+',
131
lambda match: match.group(0).encode('unicode_escape'),
102
# performance tuning for elementree's serialiser. This should be
103
# sent upstream - RBC 20060523.
104
# the functions here are patched into elementtree at runtime.
106
escape_re = re.compile("[&'\"<>]")
109
"'":"'", # FIXME: overkill
114
def _escape_replace(match, map=escape_map):
115
return map[match.group()]
117
def _escape_attrib(text, encoding=None, replace=None):
118
# escape attribute value
122
text = elementtree.ElementTree._encode(text, encoding)
124
return elementtree.ElementTree._encode_entity(text)
126
return escape_re.sub(_escape_replace, text)
128
text = replace(text, "&", "&")
129
text = replace(text, "'", "'") # FIXME: overkill
130
text = replace(text, "\"", """)
131
text = replace(text, "<", "<")
132
text = replace(text, ">", ">")
134
except (TypeError, AttributeError):
135
elementtree.ElementTree._raise_serialization_error(text)
137
elementtree.ElementTree._escape_attrib = _escape_attrib
139
escape_cdata_re = re.compile("[&<>]")
145
def _escape_cdata_replace(match, map=escape_cdata_map):
146
return map[match.group()]
148
def _escape_cdata(text, encoding=None, replace=None):
149
# escape character data
153
text = elementtree.ElementTree._encode(text, encoding)
155
return elementtree.ElementTree._encode_entity(text)
157
return escape_cdata_re.sub(_escape_cdata_replace, text)
159
text = replace(text, "&", "&")
160
text = replace(text, "<", "<")
161
text = replace(text, ">", ">")
163
except (TypeError, AttributeError):
164
elementtree.ElementTree._raise_serialization_error(text)
166
elementtree.ElementTree._escape_cdata = _escape_cdata
169
class SerializerRegistry(registry.Registry):
170
"""Registry for serializer objects"""
173
format_registry = SerializerRegistry()
174
format_registry.register_lazy('4', 'bzrlib.xml4', 'serializer_v4')
175
format_registry.register_lazy('5', 'bzrlib.xml5', 'serializer_v5')
176
format_registry.register_lazy('6', 'bzrlib.xml6', 'serializer_v6')
177
format_registry.register_lazy('7', 'bzrlib.xml7', 'serializer_v7')